Stereology with Artificial Intelligence for Examining Cells and Circuits in Neuroscience Research

体视学与人工智能在神经科学研究中检查细胞和电路

• 批准号: RTI-2022-00548
• 负责人: SwiftGallant, Ashlyn
• 金额: $ 8.72万
• 依托单位: Memorial University of Newfoundland
• 依托单位国家: 加拿大
• 项目类别: Research Tools and Instruments
• 财政年份: 2021
• 资助国家: 加拿大
• 起止时间: 2021-01-01 至 2022-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=743080
• 关键词: Stereology; Artificial; Intelligence; Examining; Cells

Stereology is the gold-standard method for analyzing the number, shape and size of brain cells for analyzing neural circuits. It is an essential tool for any neuroscience laboratory, and is the mode through which ground-breaking scientific discoveries about the brain and neural circuits are made. Memorial University of Newfoundland (MUN) does not have any stereology equipment. With a stereology system, our researchers and students would be able to amplify their existing research questions to better understand a range of natural processes of the brain, including how specific neural circuits work and what causes them to malfunction throughout life, how sex differences in the brain manifest and how stress responses lead to neural changes. The requested stereology system with artificial intelligence (AI) automatically quantifies the number, shape and size of three-dimensional (3D) structures. This new automated deep-learning AI software from SRC Stereologer comes with the capability of detecting two cell types (i.e., NeuN - neuron stain, and Iba-1 - microglia stain), with further possibility of learning. This automatization approach uses deep learning AI in a 2-step process: first, the network "learns" the desired cells of interest by processing ground truth via images marked to show the cells of interest. Then, this technology uses this information to calculate the total number of cells of interest using the optical fractionator and stereological approach. For MUN researchers in neuroscience and related fields, access to stereology with AI will enable us to collect high-quality data on an internationally reputable scale in a time- and cost-efficient manner. With recent developments in the AI module (product of National Institute of Health initiative), cell counting is 30-60 times faster and far more accurate compared to the labor-intensive and error-prone manual method that is currently our only option. Since MUN is the only university of the island of Newfoundland, there is no practical alternative for MUN researchers to access stereology equipment. The requested equipment will allow for the growth of dozens of NSERC-funded research programs at MUN and the training of ~200 HQP each year. This stereology unit will accelerate productivity and discovery, while allowing our researchers and students to stay internationally competitive in our pursuit of advancing knowledge of the nervous system.

立体学是分析神经回路中脑细胞的数量、形状和大小的黄金标准方法。它是任何神经科学实验室必不可少的工具，也是关于大脑和神经回路的突破性科学发现的模式。纽芬兰纪念大学（MUN）没有任何立体设备。有了立体系统，我们的研究人员和学生将能够扩大他们现有的研究问题，以更好地理解大脑的一系列自然过程，包括特定的神经回路是如何工作的，是什么导致它们在一生中出现故障，大脑中的性别差异是如何表现的，以及压力反应是如何导致神经变化的。所要求的具有人工智能（AI）的立体系统可以自动量化三维（3D）结构的数量、形状和大小。这款来自SRC Stereologer的新型自动化深度学习人工智能软件具有检测两种细胞类型（即NeuN -神经元染色和Iba-1 -小胶质细胞染色）的能力，具有进一步学习的可能性。这种自动化方法在两步过程中使用深度学习人工智能：首先，网络通过标记显示感兴趣的细胞的图像处理地面真相来“学习”所需的感兴趣的细胞。然后，该技术利用这些信息，利用光学分馏器和立体学方法计算感兴趣的细胞总数。对于神经科学和相关领域的模联研究人员来说，通过人工智能获得立体学将使我们能够以时间和成本效益的方式在国际上享有声誉的规模上收集高质量的数据。随着人工智能模块（美国国立卫生研究院倡议的产品）的最新发展，与目前我们唯一的选择——劳动密集型、容易出错的手动方法相比，细胞计数速度快了30-60倍，而且准确得多。由于模联是纽芬兰岛唯一的一所大学，模联研究人员没有实际的替代方法来访问立体学设备。所要求的设备将允许在模联开展数十个nserc资助的研究项目，并每年培训约200名HQP。这个立体学单位将加速生产力和发现，同时使我们的研究人员和学生在追求神经系统知识的进步方面保持国际竞争力。